Incentive-oriented power‑carbon emissions trading-tradable green certificate integrated market mechanisms using multi-agent deep reinforcement learning

Due to the stricter emission reduction and renewable energy consumption goals in China, the carbon emissions trading (CET) market and tradable green certificate (TGC) market, as key means to achieve these goals, urgently need to adjust the current operating mode timely and maximize market effectiveness by introducing adaptive incentive mechanisms. Therefore, taking the power industry as an example, an incentive-oriented power-CET-TGC integrated market simulation framework based on multi-agent deep reinforcement learning is constructed. The impact of consignment auction mechanism, traditional auction mechanism, and voluntary TGC trading mechanism on the transaction situation of each market is prospectively analyzed. The results demonstrate that with the introduction of various incentive mechanisms, the transaction scale and prices of CET and TGC markets increase, and the effectiveness of carbon reduction and renewable energy consumption is significant. Among them, auction mechanisms are valid way to promote the explicit cost of carbon emissions, and the voluntary TGC trading mechanism is an important means to enhance the clean value of green energy. In addition, it is recommended to adopt consignment auction mechanism under the short-term goal “carbon peak” of pursuing economic benefits and stable emission reduction, while traditional auction mechanisms are recommended under the long-term goal of “carbon neutrality” of urgently requiring large-scale emission reduction. Moreover, the voluntary TGC trading mechanism can achieve synergy and connection between the TGC market and the green electricity market to a certain extent. These insights can provide mechanism reference for the sustainable development and construction of CET and TGC markets.